agents.memory.core.classifier

Memory classification system using LLM-based analysis.

This module provides intelligent classification of memories into types, importance scoring, and metadata extraction using language models.

Classes

MemoryClassifier

LLM-based memory classifier for automatic memory type detection and metadata extraction.

MemoryClassifierConfig

Configuration for memory classification system.

Module Contents

class agents.memory.core.classifier.MemoryClassifier(config=None)

LLM-based memory classifier for automatic memory type detection and metadata extraction.

This classifier analyzes memory content to: - Determine memory types (semantic, episodic, procedural, etc.) - Calculate importance scores - Extract entities, topics, and sentiment - Provide classification reasoning

Initialize memory classifier with configuration.

Parameters:

config (MemoryClassifierConfig)

batch_classify(contents, contexts=None)

Classify multiple memories in batch for efficiency.

Parameters:

  • contents (list[str]) – List of memory contents to classify

  • contexts (list[dict[str, Any]] | None) – Optional list of contexts for each memory

Returns:

List of MemoryClassificationResult for each content

Return type:

list[haive.agents.memory.core.types.MemoryClassificationResult]

classify_memory(content, user_context=None, conversation_context=None)

Classify a single memory content into types and extract metadata.

Parameters:

  • content (str) – Memory content to classify

  • user_context (dict[str, Any] | None) – Optional user context for classification

  • conversation_context (dict[str, Any] | None) – Optional conversation context

Returns:

MemoryClassificationResult with types, importance, and metadata

Return type:

haive.agents.memory.core.types.MemoryClassificationResult

classify_query_intent(query)

Analyze user query to determine memory retrieval intent.

Parameters:

query (str) – User query to analyze

Returns:

MemoryQueryIntent with retrieval strategy and parameters

Return type:

haive.agents.memory.core.types.MemoryQueryIntent

create_memory_entry(content, user_context=None, conversation_context=None, namespace=None)

Create a complete memory entry with automatic classification.

Parameters:

  • content (str) – Memory content

  • user_context (dict[str, Any] | None) – Optional user context

  • conversation_context (dict[str, Any] | None) – Optional conversation context

  • namespace (str | None) – Optional memory namespace

Returns:

MemoryEntry with full classification and metadata

Return type:

haive.agents.memory.core.types.MemoryEntry

class agents.memory.core.classifier.MemoryClassifierConfig(/, **data)

Bases: pydantic.BaseModel

Configuration for memory classification system.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Parameters:

data (Any)